Re: [Boost-users] question: boost::array does not have compile time checked access

From: Howard Hinnant (hinnant_at_[hidden])
Date: 2005-08-04 08:20:46

On Aug 4, 2005, at 3:51 AM, gast128 wrote:

> me22 <me22.ca <at> gmail.com> writes:
>
>>
>> On 03/08/05, gast128 <gast128 <at> hotmail.com> wrote:
>>> Dear all,
>>>
>>> I was wondering why boost::array did not offer compile time checked 
>>> access
> to
>>> its contents, something like a.get<1>(), a similar interface like 
>>> tuple.
>>> Otherwise I see not much enhancement compared to an ordinary 
>>> std::vector.
>>>
>>> Wkr,
>>> me
>>>
>>
>> boost::array is allocated on the stack like a normal array, while a
>> std::vector needs to use an allocator, usually putting it on the heap.
>>
>> Also, the normal usage pattern for arrays ( dynamic or otherwise ) is
>> looping through them, the accesses in which cannot be compile-time
>> checked.  Note, however, that operator[] does assert that the index is
>> valid--which quickly caught me a bug a little while back.
>>
>> - Scott McMurray
>
> Yes I can see for myself that there are differences between 
> boost::array and
> the vector. My point is only that one can add a compile time safe 
> accessor to
> this array. This does not invalidate the need for a run time accessor 
> like
> operator[]. Also this would create a big advantage over the vector. 
> This heap
> and stack allocation stough could just be seen as implementation 
> details.

Fwiw, the tuple interface has been added to std::tr1::array:

http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2005/n1836.pdf

Section 6.2.2.5.

-Howard